Abstract: Accurate and stable autonomous positioning is the prerequisite for mobile robots to achieve autonomous navigation in indoor environment. Aiming at the cumulative error of visual simultaneous localization and mapping (SLAM) in indoor positioning and environmental factors that cause ultra wideband (UWB) positioning accuracy to decline, this paper proposes a SLAM/UWB-based indoor fusion positioning algorithm. First of all, the algorithm is based on the extended Kalman filter (EKF), fusing the UWB global positioning coordinates and the visual SLAM displacement increment. Then considering that the measurement noise is easily affected by the complex environment, threshold detection and adaptive measurement noise estimator are introduced to suppress the influence of abnormal values and time-varying measurement noise on the performance of the filter. Finally, an intelligent mobile car is used to conduct experiments in different indoor venues. Experiments show that the algorithm is better than a single UWB or visual SLAM positioning method, and has a more stable positioning effect than the traditional extended Kalman algorithm in a complex indoor environment.